Research map / Graph learning / GNN architectures
Meta Propagation Networks for Graph Few-shot Semi-supervised Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use Meta-PN's meta-learned adaptive label propagation and pseudo-label training instead of the strongest baseline | node classification · Cora-ML 3-shot and 5-shot | test accuracy: 72.39±0.98 -> 74.94±0.25; 78.32±0.58 -> 79.88±0.15 | Helped |
| Use Meta-PN's meta-learned adaptive label propagation and pseudo-label training instead of the strongest baseline | node classification · CiteSeer 3-shot and 5-shot | test accuracy: 67.55±0.77 -> 70.48±0.34; 72.92±0.10 -> 74.14±0.50 | Helped |
| Use Meta-PN's meta-learned adaptive label propagation and pseudo-label training instead of the strongest baseline | node classification · PubMed 3-shot and 5-shot | test accuracy: 71.22±0.82 -> 73.25±0.77; 74.24±0.87 -> 77.78±0.92 | Helped |
| Use Meta-PN's meta-learned adaptive label propagation and pseudo-label training instead of the strongest baseline | node classification · MS-CS 3-shot and 5-shot | test accuracy: 86.65±0.42 -> 88.99±0.29; 90.30±0.66 -> 91.31±0.22 | Helped |
| Use Meta-PN rather than the strongest baseline | node classification · Cora-ML standard semi-supervised | test accuracy: 85.65±.23 -> 86.33±.36 | Helped |
| Use Meta-PN rather than the strongest baseline | node classification · CiteSeer standard semi-supervised | test accuracy: 75.73±.30 -> 77.13±.31 | Helped |
| Use Meta-PN rather than the strongest baseline | node classification · PubMed standard semi-supervised | test accuracy: 79.73±.31 -> 80.39±.53 | Helped |
| Use Meta-PN rather than the strongest baseline | node classification · MS-CS standard semi-supervised | test accuracy: 92.80±.17 -> 93.92±.17 | Helped |
| Replace meta-learned propagation weights with fixed teleport probabilities (ablation) | Meta-PN · 5-shot or 1.0% label-ratio tasks | test accuracy | Hurt |
| Increase propagation steps substantially in a standard GCN (ablation) | GCN · few-shot node classification | test accuracy | Hurt |
| Use larger propagation steps with static label propagation rather than standard GCN message passing (ablation) | GCN · few-shot node classification | test accuracy | Helped |
Closest papers by meaning
- Node Classification on Graphs with Few-Shot Novel Labels via Meta Transformed Network Embedding (NeurIPS 2020 · similarity 0.91)
- Why Propagate Alone? Parallel Use of Labels and Features on Graphs (ICLR 2022 · similarity 0.91)
- Normalize Then Propagate: Efficient Homophilous Regularization for Few-Shot Semi-Supervised Node Classification (AAAI 2025 · similarity 0.89)
- Unlocking the Potential of Black-box Pre-trained GNNs for Graph Few-shot Learning (AAAI 2025 · similarity 0.89)
- Graph Metanetworks for Processing Diverse Neural Architectures (ICLR 2024 · similarity 0.89)
- NLGT: Neighborhood-based and Label-enhanced Graph Transformer Framework for Node Classification (AAAI 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.